Can Draxlr keep the previous answer's referent the same in a follow-up?
Draxlr kept the previous answer’s customer set intact in the follow-up. It answered for the same five customers, did not ask who “they” were, and the second query rebuilt that same set before counting their last-90-day orders.
1 of 1 test case passed
Every test case in this scenario has a result.
The test case
Each test case is judged on its own: Pass, Fail, Not gradable, or Untested. The scenario result above is based on this test case.
"They" refers to the previous answer's setPassEvidence
Input 1: “Who are our top 5 customers by total spend?”
Input 2: “How many orders did they place in the last 90 days?”
Response 1 answers Input 1.
Response 2, the final response, answers Input 2.
Response 3 is the query the tool generated for Input 2; it is not part of the answer.
The top 5 all-time customers placed 85 orders in the last 90 days. Dev Raghavan: 18; Priya Raghavan: 17; Jae-won Park: 13; Rosa Delgado: 19; Wei Zhang: 18. This keeps the same top 5 customers by all-time total spend.
Transcribed from Proof 1 · Screen recording at 0:06
Every message the tool sent, in order (3)
- Response to input 1
The top customer by total spend is Dev Raghavan at $24,210.69. The top 5 are Dev Raghavan ($24,210.69), Priya Raghavan ($21,597.62), Jae-won Park ($19,007.14), Rosa Delgado ($18,210.72), and Wei Zhang ($17,875.23). This covers all time.
- Final response
The final response, printed above.
- Generated query for input 2 · not part of the answer
WITH top_customers AS (SELECT c.id AS customer_id, c.name, c.email, SUM(o.total_amount) AS total_spend FROM public.customers c JOIN public.orders o ON o.customer_id = c.id GROUP BY c.id, c.name, c.email ORDER BY total_spend DESC LIMIT 5) SELECT tc.customer_id, tc.name, tc.email, tc.total_spend, COUNT(o.id) AS orders_last_90_days FROM top_customers tc LEFT JOIN public.orders o ON o.customer_id = tc.customer_id AND o.placed_at >= NOW() - INTERVAL '90 days'
| ✓ | Expected: The follow-up should use the same customer set from turn 1, without asking again who “they” are or quietly changing the set.Found: It answered for the same five customers from turn 1, gave their 90-day order counts, and did not ask for clarification or shift the set. |
The follow-up used the same five-customer set from turn 1, then counted their 90-day orders, and it did not ask who “they” were. That keeps the referent stable, so the case passes.
Full table shown. The first result card displayed all 17 customers, even though the answer named only five.
All orders counted. The first answer counted every order, including cancelled ones, and said it covered all time.
Configuration and setup
How this tool was set up for the run and what the test needed in place. Each row is a fact from the run's records; a fact the records do not hold is left out, not guessed.
How this scenario is graded
How we decide Pass, Fail and Not gradable. The same rules apply to every tool tested on this scenario.
Each test case gets one result per tool: Pass, Fail or Not gradable. A test case we haven't run yet shows Untested. There are no partial results.
- Pass: the tool did everything the test expected, and nothing it said contradicts the correct answer.
- Fail: at least one expected behaviour clearly didn't happen; the test case's result says which one and quotes the tool.
- Not gradable: our evidence couldn't settle the outcome (for example a record we needed is missing). It is never counted as a fail, and the test case's result says what's missing.
Where this sits in the benchmark
This page is one cell of a larger study: one tool, one scenario. Only this benchmark's frame appears here.
| Level | Name | Scope |
|---|---|---|
| Benchmark | AI Database Agents → | 28 scenarios · 11 tools |
| Capability | Conversational Interaction → | |
| Scenario | The next question refers to the previous answer → | |
| Tool | Draxlr → |
Global scenario definition → · Global capability definition → · Draxlr product page →
History of this result
What has happened to this result since it was first published. Runs and grades are never overwritten: a retest or a re-grade publishes a new result and keeps the earlier one readable.
Act on this result
Nothing filed here edits the run or the grade. A challenge opens a review, and a review can produce a new run or a re-grade — which becomes the current result and leaves this one in the history.
You run the same kind of test against your own setup and get the same behaviour.
Agree →Yours behaves differently. Tell us what you got, with a screenshot if you have one.
Disagree →Something here is wrong — a reference value, a transcription, a grade.
Report an issue →Tell us what changed and we schedule a rerun of the failing test case. The old result stays as history.
Vendor notice →The same record is available as structured data through the AI Demos MCP server, with the counts, the coverage and every per-test-case reason carried as fields.
Verify the proof files
These files support this result. Open a file to inspect the original evidence.
File fingerprints (SHA-256)
A fingerprint identifies the exact file used for this result.